In this large study of 1021 patients, midazolam emerged as the preferred benzodiazepine for SE treatment in both children and adults, both pre- and in-hospital. The same applies to the increased use of levetiracetam and lacosamide in children and adults, while the overall use of phenobarbital, phenytoin, and valproate use decreased. Mortality and outcome remain stable despite changes in medication patterns.

Different studies have analyzed SE treatment but not its trends over time [8, 33,34,35,36,37,38,39,40,41,42]. Only two previous studies evaluated medication trends but focused mainly on ASMs. However, our data are more recent since the other two studies only included patients up to 2012 and 2016 [43,44,45].

This is the first study to examine the impact of the RAMPART and ESETT trials on medication patterns in SE. Another notable feature of this study is that it includes children and adults and analyzes them together and separately to identify treatment differences and similarities between them.

Status epilepticus must be treated promptly, ideally in the prehospital setting, since outcomes worsen when seizure control is not achieved within 1–2 hours after SE onset [6, 46, 47]. Overall, half of SE episodes in children and one-third of SE episodes in adults were treated prehospital in this study. However, we did not differentiate between SE episodes that started prehospital and those that occurred only during an in-hospital stay. Children treated prehospital were much more likely to be treated by a layperson than adults. This difference might reflect parents administering benzodiazepines to their children with known epilepsy during SE. In adults, a lack of a partner or caregiver has been shown to decrease rescue medication use [48], and adults presented with new-onset SE more frequently than children.

In this study, we aimed to investigate whether treatment patterns for first-line therapy changed during the study period. The publication of different studies favoring midazolam use during our study may have impacted the benzodiazepine medication behavior it found [13, 49, 50]. Notably, the RAMPART study was published in 2012, reporting numeric superiority and statistic noninferiority in efficacy and safety of IM midazolam over IV lorazepam in the prehospital setting for first-line SE medication [13]. A secondary analysis of the pediatric patients in the RAMPART study showed that the overall cohort’s results also apply to them [15]. A meta-analysis from 2010 found midazolam to be superior to diazepam in any application mode and non-parenteral administration of midazolam to be as effective as parenteral administration of diazepam in pediatric SE therapy [49].

Our results show an increase in midazolam use in pre- and in-hospital first-line treatment for adults and children. Shortly after the RAMPART study was published and after the approval of buccal midazolam in the European Union (EU) in 2011 [51], midazolam became the preferred benzodiazepine in the prehospital setting for the overall cohort in 2014–2015. Midazolam was already the most frequently used prehospital benzodiazepine in adults when this study commenced. Nevertheless, there was a trend to use it even more frequently by the end of the study. In addition, there was a trend of increasing prehospital midazolam usage in children.

A more pronounced increase in in-hospital midazolam prescription was seen during the study period. Midazolam replaced lorazepam in adults and diazepam in children as the most frequently used benzodiazepine. Our results show a significant increase in midazolam use as first-line therapy in in-hospital therapy in the overall cohort and adults and trends of increasing usage in in-hospital pediatric therapy and prehospital treatment across all age groups. Several factors may have contributed to the increase in midazolam use. First, different publications have highlighted the efficacy and safety of midazolam. Second, the training of paramedic staff in our region is based on the algorithms for emergency care in the German state of Hesse. These exclusively recommend midazolam (IM or IV) for paramedics as first-line therapy [52]. Third, midazolam administration by parents has presumably become more common after the approval of buccal midazolam in the EU since this application route might be perceived as more socially compatible than other possible application types for laypersons. While in-hospital treatment was performed at different departments in our university hospital (e.g., neurology, pediatrics, and the emergency department) that comprise many treating clinicians, its internal guidelines recommend using buccal or intranasal midazolam until an IV access becomes available [53].

While this study was single-centered, and it could be argued that a general statement on changed medication behavior cannot be made based upon it, it included the medicating behavior of many physicians. For example, prehospital therapy is performed by various emergency physicians and paramedics not employed by Frankfurt University Hospital. In-hospital treatment was performed at different departments in our hospital such as neurology, pediatrics, and the emergency department that include many healthcare providers.

We found a general decline in traditional ASM use and an increase in newer ASM use, consistent with previous studies on medication use in SE [33,34,35,36, 41, 43, 44, 54]. Phenytoin has for many years been frequently used for second-line therapy in patients with SE. However, the ESETT randomized trial published in 2019 reported no difference in efficacy and safety between fosphenytoin, levetiracetam, and valproate. Therefore, any of these three drugs may be used for second-line SE treatment. Nevertheless, levetiracetam had the highest posterior probability of being the most effective in that trial [19]. A review of recent studies concluded that levetiracetam causes less frequent and often less severe side effects than phenytoin while being equally efficient [55]. Moreover, a Pakistani randomized controlled trial reported that levetiracetam might be more efficient than phenytoin in treating convulsive SE in children [56]. In this study, phenytoin had already disappeared from the top three most frequently used ASMs by 2012–2013, and its use continued to decline over time.

Most studies on SE medication report levetiracetam as the most frequent ASM used, consistent with the findings of the ESETT trial and our results [19, 33,34,35,36,37, 43, 54, 57]. Only one recent study from Norway published in 2018 showed that phenytoin remained the leading ASM, with levetiracetam only the fourth most used [39]. Their study period was 2001–2017, and evidence for levetiracetam being equally safe and effective as phenytoin was just emerging [58, 59].

In adults, we showed a consistently high prescription rate for levetiracetam, contrasting with the results of a study from Switzerland published in 2017, that found levetiracetam use still rising steadily [43, 45]. This difference can be explained by levetiracetam only being approved in Switzerland in 2008, during the study period of the Swiss study. In contrast, our study only started in 2012. Additionally, levetiracetam was already approved in Germany in 1999, and an IV solution became available in 2006 [60]. Therefore, its use might have already been established before our study [6].

The Swiss study also reported stable valproate use, while our results show a significant decline in valproate use in the overall cohort and adults [43, 45]. Valproate use declined in our cohort since it should be avoided in women of reproductive age due to its teratogenicity [61]. We cannot confirm the findings of the Swiss study, where the reported use of newer ASMs was accompanied by an increase in patients discharged with a disability (i.e., an mRS score of 3–5) [43, 45]. In this study, the proportion of episodes with an mRS score of 3–5 at discharge remained constant over the study period. Mortality did not change significantly throughout the study period, with an average of 16.5 % of episodes being fatal at discharge, consistent with previous cohort studies on SE [9, 33,34,35,36,37, 43,44,45]. In addition, the mortality at 30 days after discharge (17.3 % in 2012–2013 and 21.8 % in 2020–2021) remained relatively stable over time. However, patients treated in 2020–2021 (mean = 55.2 ± 28.2 years, median = 60.6, range = 0.1–96.5) were older than those in 2012–2013 (mean = 45.3 ± 31.5, median = 54.7, range = 0.1–93.3), which may have negatively impacted survival.

Several studies on SE treatment [8, 62,63,64] and comparisons with the dosage recommended by the German SE guidelines for adults indicate that both ASMs and benzodiazepines were usually underdosed in bolus administration, which is consistent with our adult cohort. In contrast, most benzodiazepines were correctly dosed in our pediatric cohort. However, dosing in children is usually higher than in adults, likely attributable to the lower body weight of children. Second-line bolus therapy was generally underdosed for all ASMs in adults and children, except for lacosamide. However, no recommendations for maintenance doses exist in current SE guidelines [10].

Notably, ASM maintenance doses in adults generally align with the guideline recommendations for bolus administration in SE [10]. Therefore, we anticipate that an ASM bolus is usually initially underdosed. However, when it does not terminate the SE, a maintenance dose consistent with the guideline’s recommendations for bolus dosage is achieved during therapy.

Notably, while the German SE guidelines do not provide recommendations specifically for children, some international guidelines do include them. The recommended benzodiazepine dosages from several international agencies [11, 12, 65,66,67,68,69,70,71] and the German SE guidelines [10] are summarized separately for adults and children in Table 2. There are recommended ranges for lorazepam (0.05–0.1 mg/kg BW), midazolam (0.1–0.2 mg/kg BW), diazepam (0.15–0.5 mg/kg BW), and clonazepam (0.01–0.05 mg/kg BW). Based on these ranges, our average doses are all within those recommended, except for midazolam. However, midazolam is often not administered IV, and Table 2 refers only to IV administration. It should be noted that this study evaluated the dosages for lorazepam (three episodes) and clonazepam (eight episodes) in only a few cases in children.

Propofol was the most frequently used drug for third-line therapy, only prescribed slightly more often than continuous midazolam infusion since both drugs were usually given together. Our results are consistent with those of other SE treatment studies, most of which found propofol to be given most often [33, 37, 39], with only one study describing continuous midazolam infusion as the most used [36]. However, this study included only a small number of third-line therapy cases, especially in the pediatric cohort (196 adult episodes and 32 pediatric episodes).

This study had certain limitations. First, it was single-centered and conducted at Frankfurt University Hospital, which primarily has an urban catchment area that does not represent the entire German population. Second, we investigated diverse patients of all ages with various SE severities, comorbidities, etiologies, and semiologies. Therefore, individual patients may have already had differing prognoses before treatment. Third, the German SE guidelines and treatment recommendations were updated during our study period (in 2020), which might have influenced the medication or dosing trends we identified. Fourth, the study’s retrospective design made us reliant on the completeness and accuracy of medical records from different departments. Therefore, documentation errors are possible, and under-ascertainment of cases cannot be ruled out.